在利用部分最小二乘原理进行粗差定值定位时，模型的法方程矩阵可能存在病态性，使得到的粗差定值定位结果不可靠。文中针对观测数据包含多个粗差且法方程病态问题，利用岭估计处理病态问题，建立部分最小二乘岭估计的粗差定值定位方法，给出粗差搜索步骤，利用迭代算法实现多个粗差的定值和定位。通过模拟算例分析部分最小二乘法、部分最小二乘岭估计在粗差搜索方面的效果，从另一个角度探讨粗差处理方法，推广现有的误差理论，证明文中方法的有效性。%By using the principle of partly least square to locate and value the gross errors in the observational data , the matrix equation of the adjustment model may have ill‐posed problems , w hich results in the locations and evaluations of the gross error unreliable . For the observational data w hich contain gross errors and have ill‐posed problems ,firstly ,this paper chooses the ridge estimation to deal with the ill‐posed problems ,then proposes a method based on the partly least square ridge estimation to locate and value the gross errors ,and the specific search steps for gross errors are given .The locations and evaluations of the gross error can be gotten through the iterative algorithm .At last ,the method of partly least square and the partly least square ridge estimation to locate and value the gross errors in one simulating computation are analyzed .The proposed method’s validity is verified .This method discusses the gross error approach from another perspective ,which will extend the existing error theory .
展开▼
